When an author publishes a book, they often relinquish certain controls over how their work is perceived or utilized by readers worldwide. In today’s context, that includes AI technology. Many authors have expressed upset upon discovering their works have been used to train various AI systems, leading to contentious fights and in some cases, lawsuits. Case in point, the Authors Guild’s recent suit against AI, though there remain significant misconceptions about the nuances of copyright law.
The origin of this discussion is Books3, a composite of multiple book collections used for AI training. The Atlantic allowed public access to search which authors’ works have been included in the dataset, sparking outrage and further debate on the ethical considerations surrounding the use of these works.
In an article, Ian Bogost, whose books were found in the dataset, shared why he was content with having his books included in Books3. He stated that, despite unwarranted backlash, authorship inherently invites unpredictable utilization of one’s work – a sentiment shared with other authors, albeit privately.
Bogost also emphasized that Books3’s creation was intended to resist dominance by corporate entities, allowing open-source, grassroots AI projects to compete with larger corporations. Meanwhile, the AI model by Meta, potentially uses Books3 data and is free for research and commercial use, hinting to a more complex situation.
In conclusion, despite the apprehension, authors releasing their works should come to accept that they lose control over how their works are perceived and used – whether by readers or AI technologies. And the fundamental purpose of the Books3 database, to enhance knowledge accessibility, aligns with the ideals of open-source proponents like Aaron Swartz.
For more details on this, read the original article here.